Research

Test the claim, not just the tool

Research becomes useful to an examiner when another person can identify the inputs, repeat the method, and understand the boundaries of the result. A tool's successful execution is not a validation result. Record expected behavior before examining the output.

• Design a small forensic validation study

  An original protocol for controls, known inputs, independent checks, and reproducibility.

• Provenance and detector limits

  Separate assertions, bindings, detector scores, and a conclusion about origin.

• NIST Computer Forensics Tool Testing

  Primary test-method and tool-report resource; select a report relevant to the feature and version.

• Synthetic Forensics: provenance and authenticity

  Project-authored orientation to origin and transformation claims, with its own sources and limits.

Human interpretation and source monitoring

A witness's account and a digital artifact are different sources with different failure modes. The source-monitoring research guide discusses distinctions relevant to human-AI interaction. Synthetic Imprinting is a proposed framework about lasting influence from generated or simulated experiences. The proposal is not an established forensic finding, a clinical diagnosis, or proof that an effect occurred in a witness.

What is original here?

The study-design and interpretation guides are editorial syntheses, not reports of completed experiments. No benchmark dataset, measured error rate, or new psychological effect is claimed. A future experimental report should publish its protocol, permissible inputs, results, and revision history separately.